Primary care, HMO enrollment, and hospitalization for ambulatory care sensitive conditions: a new approach

To examine the association of primary care availability, HMO enrollment, and other person and location variables with potentially preventable hospitalization for adults in New York State, compared with other types of hospitalization.

Hospital discharges of New York residents in the age group 20-64 hospitalized either in New York or in three contiguous states: New Jersey, Pennsylvania, or Connecticut using 1995 statewide discharge files from the Health care Cost and Utilization Project (HCUP) maintained by the Agency for Health care Research and Quality.

A multinomial logit model uses the individual discharge as the unit of analysis. ACS admissions are compared with marker admissions (urgent but non-ACS) and referral sensitive surgeries (more discretionary), controlling for severity of illness.

Higher primary care density was associated with a lower likelihood of ACS admission, compared with marker admissions, without increasing referral-sensitive admissions. The study also supports the hypothesis of ACS admissions being less likely for private HMO enrollees than for other insured adults. This result was not found for Medicaid HMO enrollees, even by comparison to other Medicaid enrollees.

A key policy-relevant result is the negative association of primary care physicians per capita with the likelihood of ACS admissions, without an offsetting association with resource costs via referral-sensitive admissions. The method allowed for examining the possible effects of personal and area variables on one type of hospital admission (ACS) by contrast with other specific types of admissions. This type of analysis could be strengthened in several ways for a defined population when better data on individual patients and several time periods are used.